gms | German Medical Science

Introduction: Observational Medical Outcomes Partnership (OMOP) [1] is a research repository provided by Observational Health Data Sciences and Informatics (OHDSI) [2]. As described in Reinecke et al. [3], it is becoming a trend to use OMOP in the field of medical informatics to harmonize electronic health records (EHR) data for research networks around the globe.

The Medical Informatics in Research and Care in University Medicine (MIRACUM, FKZ 01ZZ1801A/L) [4] use case 1 (UC1) [5] focuses on the improvement of patient recruitment for clinical studies using Medical Informatics Initiative (MI-I) [6] core data set (CDS) [7] in Fast Healthcare Interoperability Resources (FHIR) format and OMOP. For this purpose, we present an approach on how to ensure unified and consistent data transfer to OMOP across multiple sites.

Method: To ensure data transfer to OMOP in a unified manner an OMOP infrastructure and extract transfer load (ETL) process are required by UC1 as technical foundation. Both deliverables must be platform independent deployable and must be transferable across different systems architectures.

To foster collaborative work as a team across sites we established a recurring feedback cycle to gather input for further improvement. Additionally we conducted workshops to assist with the OMOP infrastructure and ETL job execution.

Results: We supplied a dockerized OMOP database infrastructure based on version 5.3.1, including ATLAS, the WebAPI and a common set of terminologies that ensure semantic interoperability [8]. The ETL process [9] is dockerized as well and is compatible with the MI-I CDS version 1.0. Both deliverables have been successfully deployed at each site.

Major requirements identified by the feedback are the support of more recent OMOP versions and terminologies, and the enablement of the ETL job to operate on additional MI-I CDS extension modules. Additionally, performance issues have been mentioned multiple times that need to be addressed perspectively. Moreover, we conducted different kinds of workshops brought knowledge to all sites on how to operate the deliverables. It also ensures a steady communication between all partner sites for future collaboration on OMOP.

Discussion: The transfer of the MI-I CDS data to OMOP was successfully demonstrated within MIRACUM. Our approach paves the way for further enrollment of the deliverables across other MI-I sites outside MIRACUM and thus enables the participation in national studies on OMOP. Furthermore, extended usage of OMOP in the German research community can be reached by supporting data other than those already covered by the MI-I CDS base modules, e.g. imaging [10] and genomic data [11], for which initial steps have been started already [12].

The authors declare that they have no competing interests.

The authors declare that an ethics committee vote is not required.